Struct TestRunner 

pub struct TestRunner {
    pub verbose: bool,
    pub filter: Option<String>,
    pub config: CompilerConfig,
}

Test runner configuration

Fields

verbose: bool

Show verbose output

filter: Option<String>

Filter pattern for test names

config: CompilerConfig

Compiler configuration

Implementations

impl TestRunner

pub fn new(verbose: bool, filter: Option<String>) -> Self

pub fn discover_test_files(&self, paths: &[PathBuf]) -> Vec<PathBuf>

Discover test files in the given paths

pub fn validate_paths(&self, paths: &[PathBuf]) -> Result<(), String>

Validate explicit paths before discovery. Any path ending in .seq is treated as a file path and must match test-*.seq; directories are fine (descent already filters). The check is name-based, not filesystem-state-based, so it’s deterministic in tests and surfaces the naming rule even before the file is stat’d. Converts the previously silent zero-tests outcome on misnamed files into a loud error.

pub fn discover_test_functions( &self, source: &str, ) -> Result<(Vec<String>, Vec<SkippedTest>, bool), String>

Discover test functions in a source file.

A word is a test entry point iff its name starts with test- AND its declared stack effect is exactly ( -- ). Words matching the name pattern but with a different effect (e.g. a test-flag helper with ( Int Int -- Bool )) are reported as skipped so the runner can tell the user why the synthetic main isn’t calling them — see issue #435 for the original confusion.

Returns (test_names, skipped, has_main).

pub fn run_file(&self, path: &Path) -> FileTestResults

Run all tests in a file

pub fn run(&self, paths: &[PathBuf]) -> TestSummary

Run tests and return summary

pub fn print_results(&self, summary: &TestSummary)

Print test results